Various devices and monitoring systems have been developed and deployed in order to monitor the power grid. Indeed, several real-world cyberattacks on power grid systems have been publicly reported. For the transmission and distribution, Phasor Measurement Units (PMUs) constitute the main sensing equipment of the overall wide area monitoring and situational awareness systems by collecting high-resolution data and sending them to Phasor Data Concentrators (PDCs). In this paper, we consider data spoofing attacks against PMU networks. The data between PMUs and PDC(s) are sent through the legacy networks, which are subject to many attack scenarios under with no, or inadequate, countermeasures in protocols, such as IEEE 37.118-2. We consider one potential attack, where an adversary may simply keep injecting a repeated measurement through a compromised PMU to disrupt the monitoring system. This attack is referred to as a Repeated Last Value (RLV) attack. We develop and evaluate countermeasures against RLV attacks using a 2D Convolutional Neural Network (CNN)-based approach, which operates in frames for each second mimicking images, in order to avoid the computational overhead of the classical sample-based classification algorithms, such as SVM. Further, we take this frame-based approach and use it with Support Vector Machine (SVM) for performance evaluation. Our preliminary results show that frame-based CNN as well as SVM provide promising results for RLV attacks while the efficacy of CNN over SVM frame becomes more pronounced as the attack intensity increases.

Social media is an important source of real-world data for sentiment analysis. Hate speech detection models can be trained on data from Twitter and then utilized for content filtering and removal of posts which contain hate speech. This work proposes a new algorithm for calculating user hate speech index based on user post history. Three available datasets were merged for the purpose of acquiring Twitter posts which contained hate speech. Text preprocessing and tokenization was performed, as well as outlier removal and class balancing. The proposed algorithm was used for determining hate speech index of users who posted tweets from the dataset. The preprocessed dataset was used for training and testing multiple machine learning models: k-means clustering without and with principal component analysis, naïve Bayes, decision tree and random forest. Four different feature subsets of the dataset were used for model training and testing. Anomaly detection, data transformation and parameter tuning were used in an attempt to improve classification accuracy. The highest F1 measure was achieved by training the model using a combination of user hate speech index and other user features. The results show that the usage of user hate speech index, with or without other user features, improves the accuracy of hate speech detection.

This paper considers calculation methods for the electric field intensity and magnetic flux density in the vicinity of the overhead transmission lines, as well as the calculation of alternating current (AC) corona onset electric field intensity. Calculations within this paper are made using the 2D algorithms of Charge Simulation Method (CSM) and Biot-Savart (BS) law based method. In order to obtain more accurate results, calculations are made by representing each overhead transmission line conductor with a large number of electric and magnetic field sources. By applying this approach, bundle conductors can be represented in a more realistic way and also singularity problems can be avoided when calculating electric field intensity. The presented methods are applied to a real overhead transmission line configuration, and obtained results are compared with field measurement results over the lateral profile. For considered overhead transmission line, AC corona onset electric field intensity is calculated and compared with calculated electric field intensity on the conductor’s surface. A comparison of calculated and measured results shows that considered calculation methods give satisfactory results.

In the fields of computer vision and digital image processing, image segmentation denotes a process whereby an image is segmented into multiple regions. Image segmentation based on multilevel thresholding has received significant attention in recent literature. In this paper, a multilevel thresholding approach based on three different Rao algorithms and Kapur’s entropy is investigated. The performance of the considered thresholding methods is evaluated on a dataset of 10 standard benchmark images using the mean of objective function values, the standard deviation of objective function values, and the best objective function value obtained over a fixed number of independent runs. The experimental results demonstrate the effectiveness of the multilevel thresholding approach based on Rao algorithms and Kapur’s entropy.

The broader use of devices powered by rechargeable batteries, especially constrained embedded devices, makes the efficient Battery Management System (BMS) increasingly more important. The estimation accuracy of the amount of remaining charge in the battery is critical as it affects the device’s operation and reliability. For that reason, the estimation of state-of-charge (SoC) is considered one of the main functionalities of a BMS. However, SoC estimation remains a complex task that depends on a range of internal and external factors. Most traditional SoC estimation methods are either computationally complex, require special laboratory equipment or additional configuration efforts. In addition, most methods require continuous measurement of battery parameters, which, in turn, renders these methods not applicable to the class of constrained embedded devices. This paper aims to extend the Coulomb counting method to the class of duty-cycled energy-constrained devices by designing an algorithm that combines voltage-based evaluation and pre-recorded task power profiles to estimate the SoC. In addition, a setup for identifying the battery parameters and algorithm validation setup were also developed and described in the paper.

Ambient conditions, especially temperature and humidity, have a huge impact on the performance of an air quality sensor. In this paper, four correction models were built to compensate the impact of ambient conditions. Linear regression and machine learning algorithms were used for building the models. Correction models were trained by using three types of measurement data. Raw measurement data was used in the first case. Secondly, measurement data was corrected and a significant improvement was shown. Lastly, measurements of various ambient conditions were used as well. Using corrected and extended measurement data brought a great improvement in accuracy of the models. A neural network correction model proved to be the most efficient in all cases. Compensating the impact of ambient conditions on the performance of an air quality sensor by using correction models was efficient and this method could be used in the air quality monitoring applications. This is of particular importance for usage of low-cost sensors in the air quality monitoring.

New generation networks are facing ever greater demands. When testing new network devices that must process packets at extremely high rates, it is essential to test their functionality and desired performance under maximum traffic load. As a result, in order to test the hardware, a traffic generator is required. This paper proposes an affordable and extensible high-speed FPGA-based Ethernet traffic generator. The proposed solution is able of fully utilizing a 40GbE link, with the possibility of manipulating traffic characteristics at the level of an individual packet. Although intended to run on the DE10-Pro system, the proposed design is portable to other FPGA boards with minimal development effort and changes.

A QKD network can be considered an add-on technology to a standard communication network that provides IT-secure cryptographic keys as a service. As a result, security challenges resulting in the suspension of functional work must be addressed. This study analyzes a Denial of Service (DoS) attack on the Key Management System (KMS), one of the critical components of the QKD network in charge of key management and key provisioning to authorized consumers. Through simulation methods performed in the QKDNetSim, we show that legitimate customers experience significantly worse service during an excessive DoS attack on KMS.

The article presents a series of measurements conducted on the fully-operated Quantum Key Distribution system. These measurements primarily focus on the Quantum Bit Error Rate (QBER), which is the most important parameter of the quantum channel. This parameter was observed and measured for 16 days under the quantum channel’s operating conditions to determine any correlations between the QBER and other quantum link parameters, such as secret key rate. A thorough statistical analysis of the measured data was performed as a part of this investigation and is presented in the paper.

This paper treats the problem of 3D outdoor environment mapping using images acquired by Unmanned Aerial Vehicle (UAV). The main focus is on the generation of 3D model for large scale environments. In order to perform 3D model reconstruction and mapping from 2D aerial images we employed a Structure from Motion (SfM) based approach. The obtained results using this approach for different scenarios, the rubble field and village, are presented. The generated UAV 3D point cloud data are compared with the ground truth using the least square method, where the ground truth represents a reference model with high accuracy geodetic precision. The comparison of the 3D environment models with the rubble field and village scenarios and the ground truth data is also given.

The Advanced Distribution Management System (ADMS) has grown to be a highly complicated system that comprises distribution generation, batteries, power electronics, and, in case of an urban area, an electric transportation system. One of the most essential features of ADMS is maintaining node voltages and branch thermal ratings within defined limits while maintaining minimal system losses and maximizing the use of renewable energy. Voltage VAr control (VVC) is extensively used to address these challenges and is becoming increasingly significant in ADMS. A side from the necessity to manage the system status, VVC must be adaptable to accommodate future Smart City (SC) requirements such as electric-vehicle charging and energy recuperation management. The majority of existing systems control the DC electric transportation system separately from the entire AC system. This paper attempts to tackle the problem using a hybrid single model that incorporates both: AC and DC network components.

The scientific discipline of Computer Vision (CV) is a fast developing branch of Machine Learning (ML). It addresses various tasks important for robotics, medicine, autonomous driving, surveillance, security or scene understanding. The development of sensor technologies enabled wide usage of 3D sensors, and therefore, it increased the interest of the CV research community in creating methods for 3D sensor data. This paper outlines seven CV tasks with 3D point cloud data, state-of-the-art techniques, and datasets. Additionally, we identify key challenges.

Background: Rising expenditure for new cancer medicines is accelerating concerns that their costs will become unsustainable for universal healthcare access. Moreover, early market access of new oncology medicines lacking appropriate clinical evaluation generates uncertainty over their cost-effectiveness and increases expenditure for unknown health gain. Patient-level data can complement clinical trials and generate better evidence on the effectiveness, safety and outcomes of these new medicines in routine care. This can support policy decisions including funding. Consequently, there is a need for improving datasets for establishing real-world outcomes of newly launched oncology medicines. Aim: To outline the types of available datasets for collecting patient-level data for oncology among different European countries. Additionally, to highlight concerns regarding the use and availability of such data from a health authority perspective as well as possibilities for cross-national collaboration to improve data collection and inform decision-making. Methods: A mixed methods approach was undertaken through a cross-sectional questionnaire followed-up by a focus group discussion. Participants were selected by purposive sampling to represent stakeholders across different European countries and healthcare settings. Descriptive statistics were used to analyze quantifiable questions, whilst content analysis was employed for open-ended questions. Results: 25 respondents across 18 European countries provided their insights on the types of datasets collecting oncology data, including hospital records, cancer, prescription and medicine registers. The most available is expenditure data whilst data concerning effectiveness, safety and outcomes is less available, and there are concerns with data validity. A major constraint to data collection is the lack of comprehensive registries and limited data on effectiveness, safety and outcomes of new medicines. Data ownership limits data accessibility as well as possibilities for linkage, and data collection is time-consuming, necessitating dedicated staff and better systems to facilitate the process. Cross-national collaboration is challenging but the engagement of multiple stakeholders is a key step to reach common goals through research. Conclusion: This study acts as a starting point for future research on patient-level databases for oncology across Europe. Future recommendations will require continued engagement in research, building on current initiatives and involving multiple stakeholders to establish guidelines and commitments for transparency and data sharing.